Add aggregations over distinct inputs to AggregationFuzzer #8328

duanmeng · 2024-01-10T14:42:12Z

Enhance AggregationFuzzer to add query plans that run aggregations over
unique inputs, i.e. SELECT k1, k2, count(distinct x) FROM t GROUP BY 1, 2.

The distinct column must be a comparable type for distinct aggregation.
Like Sorted aggregations, Distinct aggregations are tested using 2 logically
equivalent plans: Aggregation over Values and TableScan.

Distinct aggregations cannot be split into partial and final, hence, such
plans are not tested.

Exclude approx_xxx aggregations since their verifiers may not be able
to verify the results. The approx_percentile verifier would discard
the distinct property when calculating the expected result, say the
the expected result of the verifier would be approx_percentile(x), which
maybe different from the actual result of approx_percentile(distinct x),
when the x has lots of duplicated items.

Part of #7666

netlify · 2024-01-10T14:42:18Z

✅ Deploy Preview for meta-velox canceled.

Name	Link
🔨 Latest commit	`f205c2e`
🔍 Latest deploy log	https://app.netlify.com/sites/meta-velox/deploys/65ab7a975c36d90008cf0f33

mbasmanova

@duanmeng Meng, thank you for working on this extension to the Fuzzer. Some initial questions. Have you found any bugs using this extension?

velox/exec/fuzzer/AggregationFuzzer.cpp

duanmeng · 2024-01-11T01:16:41Z

@mbasmanova Hi Masha, I've run for 3600s using this PR's version and changed the ratio of distinct aggregation testing to 90% (vectorFuzzer_.coinToss(0.9)) locally and found no error.

I20240111 00:39:32.758095 907797 AggregationFuzzer.cpp:456] ==============================> Done with iteration 326
I20240111 00:39:32.758114 907797 AggregationFuzzer.cpp:1148] Total functions tested: 7
I20240111 00:39:32.758121 907797 AggregationFuzzer.cpp:1149] Total masked aggregations: 5 (1.53%)
I20240111 00:39:32.758127 907797 AggregationFuzzer.cpp:1151] Total global aggregations: 2 (0.61%)
I20240111 00:39:32.758131 907797 AggregationFuzzer.cpp:1153] Total group-by aggregations: 31 (9.48%)
I20240111 00:39:32.758136 907797 AggregationFuzzer.cpp:1155] Total distinct aggregations: 292 (89.30%)
I20240111 00:39:32.758141 907797 AggregationFuzzer.cpp:1157] Total aggregations over sorted inputs: 7 (2.14%)
I20240111 00:39:32.758144 907797 AggregationFuzzer.cpp:1159] Total window expressions: 2 (0.61%)
I20240111 00:39:32.758148 907797 AggregationFuzzer.cpp:1161] Total aggregations verified against reference DB: 132 (40.37%)
I20240111 00:39:32.758152 907797 AggregationFuzzer.cpp:1163] Total aggregations not verified (non-deterministic function / not supported by reference DB / reference DB failed): 318 (97.25%) / 407 (124.46%) / 52 (15.90%)
I20240111 00:39:32.758157 907797 AggregationFuzzer.cpp:1168] Total failed aggregations: 3 (0.92%)
[==========] Running 0 tests from 0 test suites.
[==========] 0 tests from 0 test suites ran. (0 ms total)
[  PASSED  ] 0 tests.

duanmeng · 2024-01-11T01:27:56Z

Have you found any bugs using this extension?

@mbasmanova Yes, I found some issues during the implementation and testing.

DistinctAggregations do not support VARBINARY. I fixed it in this PR.
The test plan with 2 drivers may fail with the following error when the input is TableScan , so change the driver number to 1.

-- Aggregation[SINGLE [g0, g1] a0 := count(ROW["c0"]) distinct] -> g0:ARRAY<BIGINT>, g1:MAP<VARBINARY,SMALLINT>, a0:BIGINT
  -- TableScan[table: hive_table] -> g0:ARRAY<BIGINT>, g1:MAP<VARBINARY,SMALLINT>, c0:INTEGER, c1:INTEGER
I20240110 17:03:27.754366 698605 Task.cpp:1082] All drivers (2) finished for task test_cursor 22 after running for 44 ms.
I20240110 17:03:27.754391 698605 Task.cpp:1766] Terminating task test_cursor 22 with state Finished after running for 44 ms.
/Users/macduan/work/native/velox-8/velox/exec/tests/utils/QueryAssertions.cpp:1104: Failure
Failed
Expected 897, got 898
2 extra rows, 1 missing rows
2 of extra rows:
        [] | [] | 2
        [] | [] | 2

1 of missing rows:
        [] | [] | 4

The count distinct result may be wrong after running for a long time. I changed the payload types to Orderable types.

mbasmanova · 2024-01-11T01:30:36Z

@duanmeng Thank you for sharing additional context. Would be nice to include that in the PR description.

mbasmanova · 2024-01-11T01:31:34Z

Interestingly CI failure is in the fuzzer:

I20240110 16:31:48.236256 220943 AggregationFuzzerBase.cpp:377] Executing query plan: 
-- Aggregation[FINAL STREAMING [g0] ] -> g0:VARBINARY
  -- LocalMerge[g0 ASC NULLS LAST] -> g0:VARBINARY
    -- Aggregation[PARTIAL STREAMING [g0] ] -> g0:VARBINARY
      -- OrderBy[PARTIAL g0 ASC NULLS LAST] -> g0:VARBINARY, c0:INTERVAL DAY TO SECOND
        -- TableScan[table: hive_table] -> g0:VARBINARY, c0:INTERVAL DAY TO SECOND
I20240110 16:31:48.254503 221499 Task.cpp:1082] All drivers (3) finished for task test_cursor 201 after running for 17 ms.
I20240110 16:31:48.254560 221499 Task.cpp:1766] Terminating task test_cursor 201 with state Finished after running for 17 ms.
I20240110 16:31:48.265017 220943 TempDirectoryPath.cpp:31] TempDirectoryPath:: removing all files from/tmp/velox_test_tBS6zg
terminate called after throwing an instance of 'std::out_of_range'
  what():  _Map_base::at
*** Aborted at 1704904308 (Unix time, try 'date -d @1704904308') ***
*** Signal 6 (SIGABRT) (0x35f0f) received by PID 220943 (pthread TID 0x7fd7e6b594c0) (linux TID 220943) (maybe from PID 220943, UID 0) (code: -6), stack trace: ***
(error retrieving stack trace)
/bin/bash: line 8: 220943 Aborted                 (core dumped) _build/debug/velox/functions/sparksql/fuzzer/spark_aggregation_fuzzer_test --seed ${RANDOM} --duration_sec 60 --logtostderr=1 --minloglevel=0 --repro_persist_path=/tmp/spark_aggregate_fuzzer_repro

mbasmanova · 2024-01-11T01:51:52Z

The count distinct result may be wrong after running for a long time. I changed the payload types to Orderable types.

Do these need to be orderable or would it be sufficient to have them comparable?

Sounds like DistinctAggregations are missing a check for types of distinct inputs. We should get a clear failure when applying distinct to non-comparable types.

duanmeng · 2024-01-11T02:48:11Z

The count distinct result may be wrong after running for a long time. I changed the payload types to Orderable types.

Do these need to be orderable or would it be sufficient to have them comparable?

Sounds like DistinctAggregations are missing a check for types of distinct inputs. We should get a clear failure when applying distinct to non-comparable types.

@mbasmanova I've updated the PR, and rerun the fuzzer for a while and the count distinct issue is not reproduced.

removed the hard-coding array_agg verifier
used the geenrateKeys API to generate payload columns
added a comparable check in DistinctAggregations

mbasmanova · 2024-01-11T02:56:22Z

The test plan with 2 drivers may fail with the following error when the
input is TableScan, so change the driver number to 1.

Distinct like Sorted aggregations must run single-threaded. Is this what you observed? The PR description as written sounds like you don't know why the fuzzer fails, so you just change number of drivers to 1.

velox/exec/DistinctAggregations.cpp

velox/exec/fuzzer/AggregationFuzzer.cpp

duanmeng · 2024-01-11T03:11:12Z

The test plan with 2 drivers may fail with the following error when the
input is TableScan, so change the driver number to 1.

Distinct like Sorted aggregations must run single-threaded. Is this what you observed? The PR description as written sounds like you don't know why the fuzzer fails, so you just change number of drivers to 1.

@mbasmanova Looks like it does not merge the two drivers' results? I will keep resolving this first.

duanmeng · 2024-01-11T04:37:02Z

Distinct like Sorted aggregations must run single-threaded.

@mbasmanova Hi Masha, I think I know the root cause. In the fuzzer if we only use singleAggregation with multiple drivers, the AssertQueryBuilder::copyResults just copies results from all the cursors, hence for multiple drivers we should use partial & final. Distinct aggregations cannot be split into partial and final, we could only use single-threaded like sorted aggregations.

mbasmanova · 2024-01-11T12:00:48Z

@duanmeng Like I mentioned earlier, similar to sorted aggregations, distinct aggregation must run single-threaded or data must be partitioned on group-by keys among threads. You can achieve that by either setting maxDrivers = 1 or by adding local partition on grouping keys before the aggregation.

mbasmanova · 2024-01-11T12:02:14Z

TableScan plans are tested with and without spilling.

Distinct aggregations do not support spilling yet. See #7791

velox/exec/fuzzer/AggregationFuzzerBase.cpp

duanmeng · 2024-01-11T12:32:00Z

@duanmeng Like I mentioned earlier, similar to sorted aggregations, distinct aggregation must run single-threaded or data must be partitioned on group-by keys among threads. You can achieve that by either setting maxDrivers = 1 or by adding local partition on grouping keys before the aggregation.

@mbasmanova Hi Masha, thanks for your guidance, I've used the maxDrivers =1 way in this refactored version, and I will try the local partition locally too.

velox/exec/fuzzer/AggregationFuzzer.cpp

mbasmanova · 2024-01-11T13:09:08Z

velox/exec/fuzzer/AggregationFuzzer.cpp

-        auto call = makeFunctionCall(signature.name, argNames, sortedInputs);
+        // Exclude approx_xxx, merge aggregations since it conflicts with
+        // distinct semantics.
+        // TODO: Add a exclude list for aggregations that cannot be supported by


Why do we need this?

velox/exec/fuzzer/AggregationFuzzer.cpp

mbasmanova · 2024-01-11T13:10:58Z

velox/exec/fuzzer/AggregationFuzzer.cpp

@@ -684,34 +730,100 @@ void resetCustomVerifiers(
  }
 }

+void initVerifier(


naming: initializeVerifiers

Is this a pure refactoring? Maybe extract refactoring of existing code into a separate PR to make reviewing easier.

Yes, I refactored this part along the the checkResults. I will restore this only left the checkResults.

mbasmanova · 2024-01-11T13:12:16Z

velox/exec/fuzzer/AggregationFuzzer.cpp

+}
+} // namespace
+
+bool AggregationFuzzer::checkResult(


naming: checkResult name suggests that this function checks some results, however, it runs plans and compares the results. Can we think of a more descriptive name?

compareEquivalentPlanResults?

This function actually executes the plans, not only compares the results. Maybe, assertEquivalentPlanResults?

velox/exec/fuzzer/AggregationFuzzer.cpp

mbasmanova

@duanmeng This is shaping up nicely. Would you share stats from running the modified fuzzer? These are printed at the end of the run.

velox/exec/fuzzer/AggregationFuzzer.cpp

kagamiori

Hi @duanmeng, thank you for extending AggregationFuzzer to cover distinct aggregations! I left a few comments.

velox/exec/fuzzer/AggregationFuzzer.cpp

duanmeng · 2024-01-18T15:39:25Z

@mbasmanova Hi Masha, I've resolved your comments with the latest codes, and ran the fuzzer locally for 360s (I will run it nightly for 3600s) the stats are as follows. Could you please help to take a look? Thanks.

I20240118 23:33:37.139185 1331215 AggregationFuzzer.cpp:475] ==============================> Done with iteration 60
I20240118 23:33:37.139217 1331215 AggregationFuzzer.cpp:1023] Total functions tested: 22
I20240118 23:33:37.139222 1331215 AggregationFuzzer.cpp:1024] Total masked aggregations: 9 (14.75%)
I20240118 23:33:37.139232 1331215 AggregationFuzzer.cpp:1026] Total global aggregations: 5 (8.20%)
I20240118 23:33:37.139236 1331215 AggregationFuzzer.cpp:1028] Total group-by aggregations: 43 (70.49%)
I20240118 23:33:37.139241 1331215 AggregationFuzzer.cpp:1030] Total distinct: 8 (13.11%)
I20240118 23:33:37.139245 1331215 AggregationFuzzer.cpp:1032] Total distinct aggregations: 4 (6.56%)
I20240118 23:33:37.139250 1331215 AggregationFuzzer.cpp:1034] Total aggregations over sorted inputs: 4 (6.56%)
I20240118 23:33:37.139253 1331215 AggregationFuzzer.cpp:1036] Total window expressions: 5 (8.20%)
I20240118 23:33:37.139257 1331215 AggregationFuzzer.cpp:1038] Total aggregations verified against reference DB: 7 (11.48%)
I20240118 23:33:37.151767 1331215 AggregationFuzzer.cpp:1040] Total aggregations not verified (non-deterministic function / not supported by reference DB / reference DB failed): 33 (54.10%) / 17 (27.87%) / 0 (0.00%)
I20240118 23:33:37.151778 1331215 AggregationFuzzer.cpp:1045] Total failed aggregations: 6 (9.84%)
[==========] Running 0 tests from 0 test suites.
[==========] 0 tests from 0 test suites ran. (0 ms total)
[  PASSED  ] 0 tests.

mbasmanova

@duanmeng Thank you for working on this. Some comments and questions.

velox/exec/fuzzer/AggregationFuzzerBase.cpp

mbasmanova · 2024-01-19T11:29:00Z

velox/exec/fuzzer/AggregationFuzzer.cpp

    size_t numDistinct{0};

+    // Number of iterations using distinct aggregation.


using aggregations over distinct inputs.

velox/exec/fuzzer/AggregationFuzzer.cpp

mbasmanova · 2024-01-19T11:30:38Z

velox/exec/fuzzer/AggregationFuzzer.cpp

+// Returns true if specified aggregate function can be applied to distinct
+// inputs.
+bool supportsDistinctInputs(const CallableSignature& signature) {
+  if (signature.args.size() != 1) {


signature.args.empty()

We should allow 2+ argument aggregate functions? Why limit to single-arg?

velox/exec/fuzzer/AggregationFuzzer.cpp

duanmeng

@mbasmanova Hi Masha, I've updated the PR according to your comments and guidance. Could you please help to take another look? Thanks.

duanmeng · 2024-01-19T15:52:19Z

velox/exec/fuzzer/AggregationFuzzer.cpp

    size_t numDistinct{0};

+    // Number of iterations using distinct aggregation.


velox/exec/fuzzer/AggregationFuzzer.cpp

duanmeng · 2024-01-19T15:52:35Z

velox/exec/fuzzer/AggregationFuzzer.cpp

+// Returns true if specified aggregate function can be applied to distinct
+// inputs.
+bool supportsDistinctInputs(const CallableSignature& signature) {
+  if (signature.args.size() != 1) {


velox/exec/fuzzer/AggregationFuzzer.cpp

velox/exec/fuzzer/AggregationFuzzerBase.cpp

velox/exec/fuzzer/AggregationFuzzer.cpp

mbasmanova

Looks great % a couple tiny comments. Thank you for adding this functionality to help improve fuzzer coverage.

mbasmanova

Thanks.

facebook-github-bot · 2024-01-19T17:04:44Z

@mbasmanova has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

mbasmanova · 2024-01-19T18:05:04Z

@duanmeng Ci is red. Would you take a look?

duanmeng · 2024-01-20T05:55:22Z

Update: CI passed after rebase.

@mbasmanova Hi Masha, there are two ci failures in S3FileSystemTest and JoinFuzzer but I could not reproduce them locally.
For S3FileSystemTest.invalidAccessKey, I think it is caused by the unstable test env, the minioServer is unstable to cause the unexpected network error message.

logs

velox_s3file_test

https://app.circleci.com/pipelines/github/facebookincubator/velox/42841/workflows/c156d45b-7b47-4acc-b5f8-77dc893387b2/jobs/296842

287: E0119 17:41:19.555821 96311 Exceptions.h:68] Line: ../../velox/connectors/hive/storage_adapters/s3fs/S3FileSystem.cpp:93, Function:initialize, Expression:  Failed to get metadata for S3 object due to: 'Network connection'. Path:'s3://dummy/foo.txt', SDK Error Type:99, HTTP Status Code:-1, S3 Service:'Unknown', Message:'curlCode: 7, Couldn't connect to server', RequestID:'', Source: RUNTIME, ErrorCode: INVALID_STATE
287: ../../velox/connectors/hive/storage_adapters/s3fs/tests/S3FileSystemTest.cpp:134: Failure
287: Value of: e.message().find("Failed to get metadata for S3 object due to: 'Access denied'. Path:'s3://dummy/foo.txt', SDK Error Type:15, HTTP Status Code:403, S3 Service:'MinIO', Message:'No response body.'") != std::string::npos
287:   Actual: false
287: Expected: true
287: Expected error message to contain 'Failed to get metadata for S3 object due to: 'Access denied'. Path:'s3://dummy/foo.txt', SDK Error Type:15, HTTP Status Code:403, S3 Service:'MinIO', Message:'No response body.'', but received 'Failed to get metadata for S3 object due to: 'Network connection'. Path:'s3://dummy/foo.txt', SDK Error Type:99, HTTP Status Code:-1, S3 Service:'Unknown', Message:'curlCode: 7, Couldn't connect to server', RequestID:'''.
287: I0119 17:41:19.557137 96311 TempDirectoryPath.cpp:31] TempDirectoryPath:: removing all files from/tmp/velox_test_S1xysT
287: [  FAILED  ] S3FileSystemTest.invalidAccessKey (27566 ms)

JoinFuzzer

https://app.circleci.com/pipelines/github/facebookincubator/velox/42841/workflows/c156d45b-7b47-4acc-b5f8-77dc893387b2/jobs/296839

I20240119 18:41:39.149451 245782 JoinFuzzer.cpp:290] Executing query plan: 
-- HashJoin[RIGHT SEMI (PROJECT) u0=t0 AND u3=t3 AND u2=t2 AND u1=t1] -> t2:INTERVAL DAY TO SECOND, match:BOOLEAN
  -- TableScan[table: hive_table] -> u0:INTERVAL DAY TO SECOND, u1:INTERVAL DAY TO SECOND, u2:INTERVAL DAY TO SECOND, u3:VARBINARY, bp4:DATE
  -- TableScan[table: hive_table] -> t0:INTERVAL DAY TO SECOND, t1:INTERVAL DAY TO SECOND, t2:INTERVAL DAY TO SECOND, t3:VARBINARY, tp4:VARCHAR, tp5:ROW<f0:BOOLEAN,f1:INTEGER>
I20240119 18:41:39.169649 249406 Task.cpp:1082] All drivers (4) finished for task test_cursor 694 after running for 18 ms.
I20240119 18:41:39.169705 249406 Task.cpp:1766] Terminating task test_cursor 694 with state Finished after running for 18 ms.
I20240119 18:41:39.170841 245782 JoinFuzzer.cpp:312] Results: [ROW ROW<t2:INTERVAL DAY TO SECOND,match:BOOLEAN>: 500 elements, no nulls]
../../velox/exec/tests/utils/QueryAssertions.cpp:1128: Failure
Failed
Expected 500, got 500
26 extra rows, 26 missing rows
10 of extra rows:
        83070177176245426 | false
        1266269431950912509 | false
        1421626354945872417 | false
        1753879513182740235 | false
        2666551404982973180 | false
        2768977648360222480 | false
        3106077618523166233 | false
        3157969022991353850 | false
        3233218116822410557 | false
        3346020967571024721 | false

10 of missing rows:
        83070177176245426 | true
        1266269431950912509 | true
        1421626354945872417 | true
        1753879513182740235 | true
        2666551404982973180 | true
        2768977648360222480 | true
        3106077618523166233 | true
        3157969022991353850 | true
        3233218116822410557 | true
        3346020967571024721 | true

Unexpected results
E20240119 18:41:39.179034 245782 Exceptions.h:68] Line: ../../velox/exec/tests/JoinFuzzer.cpp:849, Function:verify, Expression: assertEqualResults({expected}, {actual}) Logically equivalent plans produced different results, Source: RUNTIME, ErrorCode: INVALID_STATE
terminate called after throwing an instance of 'facebook::velox::VeloxRuntimeError'
  what():  Exception: VeloxRuntimeError
Error Source: RUNTIME
Error Code: INVALID_STATE
Reason: Logically equivalent plans produced different results
Retriable: False
Expression: assertEqualResults({expected}, {actual})
Function: verify
File: ../../velox/exec/tests/JoinFuzzer.cpp
Line: 849
Stack trace:
# 0  _ZN8facebook5velox7process10StackTraceC1Ei
# 1  _ZN8facebook5velox14VeloxException5State4makeIZNS1_C4EPKcmS5_St17basic_string_viewIcSt11char_traitsIcEES9_S9_S9_bNS1_4TypeES9_EUlRT_E_EESt10shared_ptrIKS2_ESA_SB_
# 2  _ZN8facebook5velox14VeloxExceptionC1EPKcmS3_St17basic_string_viewIcSt11char_traitsIcEES7_S7_S7_bNS1_4TypeES7_
# 3  _ZN8facebook5velox17VeloxRuntimeErrorC2EPKcmS3_St17basic_string_viewIcSt11char_traitsIcEES7_S7_S7_bS7_
# 4  _ZN8facebook5velox6detail14veloxCheckFailINS0_17VeloxRuntimeErrorEPKcEEvRKNS1_18VeloxCheckFailArgsET0_
# 5  _ZN8facebook5velox4exec4test12_GLOBAL__N_110JoinFuzzer6verifyENS0_4core8JoinTypeE
# 6  _ZN8facebook5velox4exec4test12_GLOBAL__N_110JoinFuzzer2goEv
# 7  _ZN8facebook5velox4exec4test10joinFuzzerEm
# 8  _ZN16JoinFuzzerRunner3runEm
# 9  main
# 10 __libc_start_main
# 11 _start

mbasmanova · 2024-01-22T14:05:24Z

@duanmeng Meng, thank you for looking into CI failures. Would you create 2 separate GitHub issues for these 2 problems (assuming there are no existing issues)?

@majetideepak Deepak, would you help look into S3FileSystemTest.invalidAccessKey failure?

facebook-github-bot · 2024-01-22T14:28:04Z

@mbasmanova merged this pull request in e38603a.

conbench-facebook · 2024-01-22T14:49:37Z

Conbench analyzed the 1 benchmark run on commit e38603a5.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details.

duanmeng requested a review from mbasmanova January 10, 2024 14:42

facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Jan 10, 2024

duanmeng force-pushed the distinctFuzzer branch 2 times, most recently from 8f23e52 to 27839ed Compare January 10, 2024 15:21

mbasmanova requested review from kgpai and kagamiori January 10, 2024 18:28

mbasmanova reviewed Jan 10, 2024

View reviewed changes

velox/exec/fuzzer/AggregationFuzzer.cpp Outdated Show resolved Hide resolved

velox/exec/fuzzer/AggregationFuzzer.cpp Outdated Show resolved Hide resolved

duanmeng force-pushed the distinctFuzzer branch from 6947a0c to 1b67f38 Compare January 11, 2024 02:40

mbasmanova reviewed Jan 11, 2024

View reviewed changes

velox/exec/DistinctAggregations.cpp Outdated Show resolved Hide resolved

mbasmanova reviewed Jan 11, 2024

View reviewed changes

velox/exec/fuzzer/AggregationFuzzer.cpp Outdated Show resolved Hide resolved

mbasmanova reviewed Jan 11, 2024

View reviewed changes

velox/exec/fuzzer/AggregationFuzzer.cpp Outdated Show resolved Hide resolved

duanmeng force-pushed the distinctFuzzer branch 6 times, most recently from bd0537a to bb36e24 Compare January 11, 2024 09:47

mbasmanova reviewed Jan 11, 2024

View reviewed changes

velox/exec/fuzzer/AggregationFuzzerBase.cpp Outdated Show resolved Hide resolved

mbasmanova reviewed Jan 11, 2024

View reviewed changes

duanmeng force-pushed the distinctFuzzer branch from 1f8b276 to f581f34 Compare January 11, 2024 15:50

mbasmanova reviewed Jan 11, 2024

View reviewed changes

kagamiori reviewed Jan 11, 2024

View reviewed changes

duanmeng force-pushed the distinctFuzzer branch 2 times, most recently from 42088f8 to d44912f Compare January 18, 2024 15:33

duanmeng force-pushed the distinctFuzzer branch from d44912f to 4e7e431 Compare January 18, 2024 15:55

mbasmanova reviewed Jan 19, 2024

View reviewed changes

mbasmanova changed the title ~~Add distinct aggregations to AggregationFuzzer~~ Add aggregations over distinct inputs to AggregationFuzzer Jan 19, 2024

duanmeng force-pushed the distinctFuzzer branch from 4e7e431 to 8edbab8 Compare January 19, 2024 16:23

duanmeng commented Jan 19, 2024

View reviewed changes

duanmeng force-pushed the distinctFuzzer branch from 8edbab8 to ff47dbc Compare January 19, 2024 16:33

mbasmanova reviewed Jan 19, 2024

View reviewed changes

velox/exec/fuzzer/AggregationFuzzer.cpp Outdated Show resolved Hide resolved

mbasmanova reviewed Jan 19, 2024

View reviewed changes

velox/exec/fuzzer/AggregationFuzzer.cpp Outdated Show resolved Hide resolved

mbasmanova reviewed Jan 19, 2024

View reviewed changes

duanmeng force-pushed the distinctFuzzer branch from ff47dbc to 6991d90 Compare January 19, 2024 16:58

mbasmanova approved these changes Jan 19, 2024

View reviewed changes

Add distinct aggregation fuzzer

f205c2e

duanmeng force-pushed the distinctFuzzer branch from 6991d90 to f205c2e Compare January 20, 2024 07:47

facebook-github-bot closed this in e38603a Jan 22, 2024

facebook-github-bot added the Merged label Jan 22, 2024

		size_t numDistinct{0};

		// Number of iterations using distinct aggregation.

Add aggregations over distinct inputs to AggregationFuzzer #8328

Add aggregations over distinct inputs to AggregationFuzzer #8328

Conversation

duanmeng commented Jan 10, 2024 • edited Loading

netlify bot commented Jan 10, 2024 • edited Loading

✅ Deploy Preview for meta-velox canceled.

mbasmanova left a comment

Choose a reason for hiding this comment

duanmeng commented Jan 11, 2024 • edited Loading

duanmeng commented Jan 11, 2024 • edited Loading

mbasmanova commented Jan 11, 2024

mbasmanova commented Jan 11, 2024

mbasmanova commented Jan 11, 2024

duanmeng commented Jan 11, 2024

mbasmanova commented Jan 11, 2024

duanmeng commented Jan 11, 2024

duanmeng commented Jan 11, 2024

mbasmanova commented Jan 11, 2024

mbasmanova commented Jan 11, 2024

duanmeng commented Jan 11, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mbasmanova left a comment

Choose a reason for hiding this comment

kagamiori left a comment

Choose a reason for hiding this comment

duanmeng commented Jan 18, 2024

mbasmanova left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

duanmeng left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mbasmanova left a comment

Choose a reason for hiding this comment

mbasmanova left a comment

Choose a reason for hiding this comment

facebook-github-bot commented Jan 19, 2024

mbasmanova commented Jan 19, 2024

duanmeng commented Jan 20, 2024 • edited Loading

logs

velox_s3file_test

JoinFuzzer

mbasmanova commented Jan 22, 2024

facebook-github-bot commented Jan 22, 2024

conbench-facebook bot commented Jan 22, 2024

duanmeng commented Jan 10, 2024 •

edited

Loading

netlify bot commented Jan 10, 2024 •

edited

Loading

duanmeng commented Jan 11, 2024 •

edited

Loading

duanmeng commented Jan 11, 2024 •

edited

Loading

duanmeng commented Jan 20, 2024 •

edited

Loading